| 摘要 |
An input video signal representative of a scene is processed by carrying out a spatial transform operation on the input video signal so as to generate a transformed video signal. Concurrently, an input key signal representing the opacity of the scene and comprising a stream of digital data words is written into a key frame buffer. A first sequence of address words is generated, and a transform operation is carried out on the first sequence of address words so as to generate a second sequence of address words. The second sequence of address words is used to read ditial data words from the key frame buffer so as to generate a first transformed key signal. The transform that was carried out on the first sequence of address words is such that the first transformed key signal is spatially transformed with respect to the input key signal in the same manner as the transformed video signal is spatially transformed with respect to the input video signal. The second sequence of address words is modified so as to generate a third sequence of address words, and the third sequence of address words is used to read digital words from the key frame buffer and generate a second transformed key signal. The transformed video signal is combined with the first and second transformed key signals. |
